HYBRID ARTIFICIAL INTELLIGENCE MODEL FOR NETWORK ANOMALY DETECTION IN UZBEKISTAN'S DIGITAL ECONOMY: AN INTEGRATION ANALYSIS OF LSTM, RANDOM FOREST, AND AUTOENCODER

Authors

  • Salimov Nurali Ramazon ugli PhD student, Department of Digital Technologies and Artificial Intelligence, Tashkent State University of Economics (State Institution) Author

Keywords:

artificial intelligence, cybersecurity, network anomaly detection, Long Short-Term Memory, Random Forest, Autoencoder, machine learning, deep learning, intrusion detection, digital economy, Uzbekistan, network security.

Abstract

The rapid development of digital technologies has significantly increased the dependence of Uzbekistan's economy on networked information systems, cloud services, electronic government platforms, digital banking, telecommunications, and online commercial services. At the same time, the expansion of digital infrastructure has increased the potential exposure to network anomalies and cyberattacks. This study proposes a hybrid artificial intelligence model for network anomaly detection that integrates Long Short-Term Memory (LSTM), Random Forest (RF), and Autoencoder (AE) algorithms. The proposed architecture combines the temporal pattern recognition capability of LSTM, the classification and feature-importance capabilities of Random Forest, and the unsupervised anomaly detection capability of Autoencoder. The main objective is to develop a multidimensional detection framework capable of identifying both known and previously unseen abnormal network behavior. The proposed methodology includes data collection, preprocessing, feature engineering, Autoencoder-based representation learning, LSTM-based temporal analysis, Random Forest classification, and hybrid risk-score calculation. The model is particularly relevant to Uzbekistan because cybercrime and technology-enabled offenses have increased substantially alongside digital transformation. Official Uzbek government information indicates that 58,800 technology-related crimes across 62 categories were registered in 2024, compared with 863 crimes across 18 categories in 2019. The proposed hybrid approach provides a conceptual foundation for developing adaptive, scalable, and proactive network security systems for Uzbekistan's banking, telecommunications, e-government, and corporate digital infrastructure.

References

1.Breiman, L. (2001). Random forests. Machine Learning, 45, 5–32.

2.Breiman, L. (2001). Random forests. Machine Learning, 45, 5–32.

3.Buczak, A. L., & Guven, E. (2016). A survey of data mining and machine learning methods for cyber security intrusion detection. IEEE Communications Surveys & Tutorials, 18(2), 1153–1176.

4.Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press.

5.Hochreiter, S., & Schmidhuber, J. (1997). Long short-term memory. Neural Computation, 9(8), 1735–1780.

6.Mushtaq, E., Zameer, A., Umer, M., & Abbasi, A. A. (2022). A two-stage intrusion detection system with auto-encoder and LSTMs. Applied Soft Computing. The study demonstrated the value of combining Autoencoder-based feature learning with LSTM classification for intrusion detection.

7.Mhamdi, L., & Isa, M. M. (2024). Securing SDN: Hybrid autoencoder-random forest for intrusion detection and attack mitigation. Journal of Network and Computer Applications.

8.Almuhanna, R., & Dardouri, S. (2025). A deep learning/machine learning approach for anomaly based network intrusion detection. Frontiers in Artificial Intelligence.

9.HAE-HRL: A network intrusion detection system utilizing a novel autoencoder and a hybrid enhanced LSTM-CNN-based residual network. (2025). Computers & Security.

10.Russell, S., & Norvig, P. (2021). Artificial Intelligence: A Modern Approach (4th ed.). Pearson.

11.Sommer, R., & Paxson, V. (2010). Outside the closed world: On using machine learning for network intrusion detection. 2010 IEEE Symposium on Security and Privacy, 305–316.

12.Yin, C., Zhu, Y., Fei, J., & He, X. (2017). A deep learning approach for intrusion detection using recurrent neural networks. IEEE Access, 5, 21954–21961.

13.Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 16, 321–357.

14.Ministry of Internal Affairs of the Republic of Uzbekistan. (2025). Creating a Safe Cyberspace by Combating Cybercrime: Uzbekistan's Experience in Preventing Cybercrime and Ensuring a Safe Cyberspace.

15.Ministry of Internal Affairs of the Republic of Uzbekistan. (2025). Information on the activities of the Cybersecurity Center and the growth of technology-related crimes in Uzbekistan.

Downloads

Published

2026-09-16